Intent IQ: The Behind-the-Scenes Identity Workhorse

Intent IQ operates as an identity resolution and attribution platform, blending deterministic and probabilistic signals to map users and households across devices while claiming to be privacy-conscious. Here’s a breakdown of how it works under the hood:

The Basics:

• Hybrid Identity Graph: Intent IQ uses a mix of deterministic data (e.g., hashed emails tied to cookies or mobile ad IDs) and probabilistic data (behavioral patterns, connection data) to create a constantly updating "identity graph" that tracks users across devices and browsers.

How They Do It:

1. Deterministic Seed + Probabilistic Layers: They start with a deterministic data core—actual, confirmed identifiers (like hashed emails) that are verified across a variety of sources. Around that core, they layer probabilistic signals to fill in gaps where deterministic data falls short.

2. Dynamic Device Graph: Instead of static ID mapping that can decay (think cookies or simple device IDs), they claim their system updates in real-time, monitoring device usage across broadband households. They’ve mapped over 150 million households in the U.S. alone.

3. IP and User-Agent Signals: Intent IQ’s system uses IP addresses, user agents, and cookie IDs, but they insist it’s not fingerprinting. The key difference they claim is resilience: even when IPs or user agents change, their machine learning system recalibrates and maintains user identity across changes, unlike traditional fingerprinting which quickly breaks.

4. Partner Data Coop: Intent IQ aggregates billions of signals per day from its partners (DSPs, SSPs, etc.). These partners share log-level bid data (RTB logs) that feed into Intent IQ’s graph. Clients can also build their own “private graphs” using Intent IQ’s infrastructure.

5. AI and Machine Learning: Layers of algorithms—behavioral models, clustering, and pattern recognition—are applied to build persistent identity across fragmented data. This allows for attribution windows as long as six months, even in Safari or other cookieless environments.

6. Privacy Positioning: While they use IP and user-agent data, Intent IQ claims they avoid traditional fingerprinting by not creating static "keys" from data points. Instead, the system adapts and updates dynamically to maintain privacy compliance while still maintaining high match rates (90%-97%).

What Makes Them Different:

• Persistent Cross-Environment Attribution: They can resolve users across mobile apps, CTV, web, and multiple devices—even when cookie data is absent or limited.

• White-Labeled Backbone: Many major ad tech companies use Intent IQ's tech behind the scenes, but white-label it under their own brands.

TL;DR:

Intent IQ is building an identity backbone with an evolving graph of users, fueled by shared partner data and real-time recalibration of device and household connections. They operate like a behind-the-scenes utility, helping clients with attribution, targeting, and bid optimization—without relying on traditional cookie-based identity or brittle fingerprinting methods.

We spoke to Aziz Rahimtoola over email, and here’s what he had to say about IntentIQ.

“Ray is a smart guy, and the tech is real.”

IntentIQ has focused on identity resolution tied to hashed email addresses—an approach that makes sense, especially now that cookies are sticking around. Aziz was complimentary of our approach as well, calling it very effective.

That said, while he did some initial digging, they don’t use IntentIQ. He’s not sure if they’ll succeed but respects Ray a lot and believes he wouldn’t be in the game if he didn’t think he could win.

Aziz described Ray as an innovator, driven by that mindset.

Just had a chat with Robert Webster, and let’s just say he’s not exactly rolling out the red carpet for IntentIQ:

"But when I did integrations last year with Magnite, PubMatic, Magnite, and Xandr, IntentIQ was not an option. But if you want, you can say I did five big integrations last year, looked extensively at ID and match partners, and no one recommended we use them. I can't speak with authority, but until I find a reason to think differently, I would put them in the time wasters category."

Time wasters. Ouch.

Become a paying subscriber to get access to this post and other subscriber-only content.

Already a paying subscriber? Sign In.